# Stockfish, a UCI chess playing engine derived from Glaurung 2.1
# Copyright (C) 2004-2008 Tord Romstad (Glaurung author)
# Copyright (C) 2008-2015 Marco Costalba, Joona Kiiski, Tord Romstad
-# Copyright (C) 2015-2016 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
+# Copyright (C) 2015-2019 Marco Costalba, Joona Kiiski, Gary Linscott, Tord Romstad
#
# Stockfish is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
### Section 1. General Configuration
### ==========================================================================
-### Establish the operating system name
-KERNEL = $(shell uname -s)
-ifeq ($(KERNEL),Linux)
- OS = $(shell uname -o)
-endif
-
### Executable name
+ifeq ($(COMP),mingw)
+EXE = stockfish.exe
+else
EXE = stockfish
+endif
### Installation dir definitions
PREFIX = /usr/local
### Built-in benchmark for pgo-builds
PGOBENCH = ./$(EXE) bench
-### Object files
-OBJS = benchmark.o bitbase.o bitboard.o endgame.o evaluate.o main.o \
- material.o misc.o movegen.o movepick.o pawns.o position.o psqt.o \
- search.o thread.o timeman.o tt.o uci.o ucioption.o syzygy/tbprobe.o
+### Source and object files
+SRCS = benchmark.cpp bitbase.cpp bitboard.cpp endgame.cpp evaluate.cpp main.cpp \
+ material.cpp misc.cpp movegen.cpp movepick.cpp pawns.cpp position.cpp psqt.cpp \
+ search.cpp thread.cpp timeman.cpp tt.cpp uci.cpp ucioption.cpp tune.cpp syzygy/tbprobe.cpp \
+ nnue/evaluate_nnue.cpp nnue/features/half_kp.cpp
+
+OBJS = $(notdir $(SRCS:.cpp=.o))
+
+VPATH = syzygy:nnue:nnue/features
+
+### Establish the operating system name
+KERNEL = $(shell uname -s)
+ifeq ($(KERNEL),Linux)
+ OS = $(shell uname -o)
+endif
### ==========================================================================
### Section 2. High-level Configuration
### ==========================================================================
#
-# flag --- Comp switch --- Description
+# flag --- Comp switch --- Description
# ----------------------------------------------------------------------------
#
# debug = yes/no --- -DNDEBUG --- Enable/Disable debug mode
-# sanitize = yes/no --- (-fsanitize ) --- enable undefined behavior checks
+# sanitize = undefined/thread/no (-fsanitize )
+# --- ( undefined ) --- enable undefined behavior checks
+# --- ( thread ) --- enable threading error checks
# optimize = yes/no --- (-O3/-fast etc.) --- Enable/Disable optimizations
# arch = (name) --- (-arch) --- Target architecture
# bits = 64/32 --- -DIS_64BIT --- 64-/32-bit operating system
# prefetch = yes/no --- -DUSE_PREFETCH --- Use prefetch asm-instruction
# popcnt = yes/no --- -DUSE_POPCNT --- Use popcnt asm-instruction
# sse = yes/no --- -msse --- Use Intel Streaming SIMD Extensions
+# ssse3 = yes/no --- -mssse3 --- Use Intel Supplemental Streaming SIMD Extensions 3
+# sse41 = yes/no --- -msse4.1 --- Use Intel Streaming SIMD Extensions 4.1
+# avx2 = yes/no --- -mavx2 --- Use Intel Advanced Vector Extensions 2
# pext = yes/no --- -DUSE_PEXT --- Use pext x86_64 asm-instruction
+# avx512 = yes/no --- -mavx512bw --- Use Intel Advanced Vector Extensions 512
+# vnni = yes/no --- -mavx512vnni --- Use Intel Vector Neural Network Instructions 512
+# neon = yes/no --- -DUSE_NEON --- Use ARM SIMD architecture
#
# Note that Makefile is space sensitive, so when adding new architectures
# or modifying existing flags, you have to make sure there are no extra spaces
# at the end of the line for flag values.
### 2.1. General and architecture defaults
+
+ifeq ($(ARCH),)
+ empty_arch = yes
+endif
+
optimize = yes
debug = no
sanitize = no
-bits = 32
+bits = 64
prefetch = no
popcnt = no
+mmx = no
sse = no
+ssse3 = no
+sse41 = no
+avx2 = no
pext = no
+avx512 = no
+vnni = no
+neon = no
+ARCH = x86-64-modern
### 2.2 Architecture specific
ifeq ($(ARCH),general-32)
arch = any
+ bits = 32
endif
ifeq ($(ARCH),x86-32-old)
arch = i386
+ bits = 32
endif
ifeq ($(ARCH),x86-32)
arch = i386
+ bits = 32
prefetch = yes
+ mmx = yes
sse = yes
endif
ifeq ($(ARCH),general-64)
arch = any
- bits = 64
endif
ifeq ($(ARCH),x86-64)
arch = x86_64
- bits = 64
prefetch = yes
sse = yes
endif
-ifeq ($(ARCH),x86-64-modern)
+ifeq ($(ARCH),x86-64-sse3-popcnt)
+ arch = x86_64
+ prefetch = yes
+ sse = yes
+ popcnt = yes
+endif
+
+ifeq ($(ARCH),x86-64-ssse3)
+ arch = x86_64
+ prefetch = yes
+ sse = yes
+ ssse3 = yes
+endif
+
+ifeq ($(ARCH),$(filter $(ARCH),x86-64-sse41-popcnt x86-64-modern))
arch = x86_64
- bits = 64
prefetch = yes
popcnt = yes
sse = yes
+ ssse3 = yes
+ sse41 = yes
+endif
+
+ifeq ($(ARCH),x86-64-avx2)
+ arch = x86_64
+ prefetch = yes
+ popcnt = yes
+ sse = yes
+ ssse3 = yes
+ sse41 = yes
+ avx2 = yes
endif
ifeq ($(ARCH),x86-64-bmi2)
arch = x86_64
- bits = 64
prefetch = yes
popcnt = yes
sse = yes
+ ssse3 = yes
+ sse41 = yes
+ avx2 = yes
+ pext = yes
+endif
+
+ifeq ($(ARCH),x86-64-avx512)
+ arch = x86_64
+ prefetch = yes
+ popcnt = yes
+ sse = yes
+ ssse3 = yes
+ sse41 = yes
+ avx2 = yes
pext = yes
+ avx512 = yes
+endif
+
+ifeq ($(ARCH),x86-64-vnni)
+ arch = x86_64
+ prefetch = yes
+ popcnt = yes
+ sse = yes
+ ssse3 = yes
+ sse41 = yes
+ avx2 = yes
+ pext = yes
+ avx512 = yes
+ vnni = yes
endif
ifeq ($(ARCH),armv7)
arch = armv7
prefetch = yes
+ bits = 32
+endif
+
+ifeq ($(ARCH),armv8)
+ arch = armv8-a
+ prefetch = yes
+ popcnt = yes
+ neon = yes
+endif
+
+ifeq ($(ARCH),apple-silicon)
+ arch = arm64
+ prefetch = yes
+ popcnt = yes
+ neon = yes
endif
ifeq ($(ARCH),ppc-32)
arch = ppc
+ bits = 32
endif
ifeq ($(ARCH),ppc-64)
arch = ppc64
- bits = 64
+ popcnt = yes
+ prefetch = yes
endif
-
### ==========================================================================
-### Section 3. Low-level configuration
+### Section 3. Low-level Configuration
### ==========================================================================
### 3.1 Selecting compiler (default = gcc)
-
-CXXFLAGS += -Wall -Wcast-qual -fno-exceptions -fno-rtti -std=c++11 $(EXTRACXXFLAGS)
-DEPENDFLAGS += -std=c++11
+CXXFLAGS += -Wall -Wcast-qual -fno-exceptions -std=c++17 $(EXTRACXXFLAGS)
+DEPENDFLAGS += -std=c++17
LDFLAGS += $(EXTRALDFLAGS)
ifeq ($(COMP),)
CXX=g++
CXXFLAGS += -pedantic -Wextra -Wshadow
- ifeq ($(ARCH),armv7)
+ ifeq ($(ARCH),$(filter $(ARCH),armv7 armv8))
ifeq ($(OS),Android)
CXXFLAGS += -m$(bits)
- LDFLAGS += -m$(bits)
+ LDFLAGS += -m$(bits)
endif
else
CXXFLAGS += -m$(bits)
ifneq ($(KERNEL),Darwin)
LDFLAGS += -Wl,--no-as-needed
endif
+
+ gccversion = $(shell $(CXX) --version)
+ gccisclang = $(findstring clang,$(gccversion))
endif
ifeq ($(COMP),mingw)
CXX=clang++
CXXFLAGS += -pedantic -Wextra -Wshadow
- ifeq ($(ARCH),armv7)
+ ifneq ($(KERNEL),Darwin)
+ ifneq ($(KERNEL),OpenBSD)
+ LDFLAGS += -latomic
+ endif
+ endif
+
+ ifeq ($(ARCH),$(filter $(ARCH),armv7 armv8))
ifeq ($(OS),Android)
CXXFLAGS += -m$(bits)
LDFLAGS += -m$(bits)
endif
ifeq ($(KERNEL),Darwin)
- CXXFLAGS += -arch $(arch) -mmacosx-version-min=10.9
- LDFLAGS += -arch $(arch) -mmacosx-version-min=10.9
+ CXXFLAGS += -arch $(arch) -mmacosx-version-min=10.14
+ LDFLAGS += -arch $(arch) -mmacosx-version-min=10.14
endif
### Travis CI script uses COMPILER to overwrite CXX
endif
### 3.2.2 Debugging with undefined behavior sanitizers
-ifeq ($(sanitize),yes)
- CXXFLAGS += -g3 -fsanitize=undefined
- LDFLAGS += -fsanitize=undefined
+ifneq ($(sanitize),no)
+ CXXFLAGS += -g3 -fsanitize=$(sanitize)
+ LDFLAGS += -fsanitize=$(sanitize)
endif
### 3.3 Optimization
CXXFLAGS += -O3
ifeq ($(comp),gcc)
-
- ifeq ($(KERNEL),Darwin)
- ifeq ($(arch),i386)
- CXXFLAGS += -mdynamic-no-pic
- endif
- ifeq ($(arch),x86_64)
- CXXFLAGS += -mdynamic-no-pic
- endif
- endif
-
ifeq ($(OS), Android)
CXXFLAGS += -fno-gcse -mthumb -march=armv7-a -mfloat-abi=softfp
endif
endif
- ifeq ($(comp),icc)
+ ifeq ($(comp),$(filter $(comp),gcc clang icc))
ifeq ($(KERNEL),Darwin)
CXXFLAGS += -mdynamic-no-pic
endif
endif
-
- ifeq ($(comp),clang)
- ifeq ($(KERNEL),Darwin)
- CXXFLAGS += -flto
- LDFLAGS += $(CXXFLAGS)
- ifeq ($(arch),i386)
- CXXFLAGS += -mdynamic-no-pic
- endif
- ifeq ($(arch),x86_64)
- CXXFLAGS += -mdynamic-no-pic
- endif
- endif
- endif
endif
### 3.4 Bits
### 3.6 popcnt
ifeq ($(popcnt),yes)
- ifeq ($(comp),icc)
+ ifeq ($(arch),$(filter $(arch),ppc64 armv8-a arm64))
+ CXXFLAGS += -DUSE_POPCNT
+ else ifeq ($(comp),icc)
CXXFLAGS += -msse3 -DUSE_POPCNT
else
CXXFLAGS += -msse3 -mpopcnt -DUSE_POPCNT
endif
endif
+ifeq ($(avx2),yes)
+ CXXFLAGS += -DUSE_AVX2
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ CXXFLAGS += -mavx2
+ endif
+endif
+
+ifeq ($(avx512),yes)
+ CXXFLAGS += -DUSE_AVX512
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ CXXFLAGS += -mavx512f -mavx512bw
+ endif
+endif
+
+ifeq ($(vnni),yes)
+ CXXFLAGS += -DUSE_VNNI
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ CXXFLAGS += -mavx512vnni -mavx512dq -mavx512vl
+ endif
+endif
+
+ifeq ($(sse41),yes)
+ CXXFLAGS += -DUSE_SSE41
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ CXXFLAGS += -msse4.1
+ endif
+endif
+
+ifeq ($(ssse3),yes)
+ CXXFLAGS += -DUSE_SSSE3
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ CXXFLAGS += -mssse3
+ endif
+endif
+
+ifeq ($(mmx),yes)
+ CXXFLAGS += -DUSE_MMX
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ CXXFLAGS += -mmmx
+ endif
+endif
+
+ifeq ($(neon),yes)
+ CXXFLAGS += -DUSE_NEON
+endif
+
+ifeq ($(arch),x86_64)
+ CXXFLAGS += -msse2 -DUSE_SSE2
+endif
+
### 3.7 pext
ifeq ($(pext),yes)
CXXFLAGS += -DUSE_PEXT
endif
endif
-### 3.8 Link Time Optimization, it works since gcc 4.5 but not on mingw under Windows.
+### 3.8 Link Time Optimization
### This is a mix of compile and link time options because the lto link phase
### needs access to the optimization flags.
-ifeq ($(comp),gcc)
- ifeq ($(optimize),yes)
- ifeq ($(debug),no)
+ifeq ($(optimize),yes)
+ifeq ($(debug), no)
+ ifeq ($(comp),clang)
+ CXXFLAGS += -flto=thin
+ LDFLAGS += $(CXXFLAGS)
+
+# GCC and CLANG use different methods for parallelizing LTO and CLANG pretends to be
+# GCC on some systems.
+ else ifeq ($(comp),gcc)
+ ifeq ($(gccisclang),)
CXXFLAGS += -flto
+ LDFLAGS += $(CXXFLAGS) -flto=jobserver
+ ifneq ($(findstring MINGW,$(KERNEL)),)
+ LDFLAGS += -save-temps
+ else ifneq ($(findstring MSYS,$(KERNEL)),)
+ LDFLAGS += -save-temps
+ endif
+ else
+ CXXFLAGS += -flto=thin
LDFLAGS += $(CXXFLAGS)
endif
- endif
-endif
-ifeq ($(comp),mingw)
+# To use LTO and static linking on windows, the tool chain requires a recent gcc:
+# gcc version 10.1 in msys2 or TDM-GCC version 9.2 are know to work, older might not.
+# So, only enable it for a cross from Linux by default.
+ else ifeq ($(comp),mingw)
ifeq ($(KERNEL),Linux)
- ifeq ($(optimize),yes)
- ifeq ($(debug),no)
CXXFLAGS += -flto
- LDFLAGS += $(CXXFLAGS)
- endif
+ LDFLAGS += $(CXXFLAGS) -flto=jobserver
endif
endif
endif
+endif
### 3.9 Android 5 can only run position independent executables. Note that this
### breaks Android 4.0 and earlier.
LDFLAGS += -fPIE -pie
endif
-
### ==========================================================================
-### Section 4. Public targets
+### Section 4. Public Targets
### ==========================================================================
help:
@echo ""
@echo "Supported targets:"
@echo ""
+ @echo "help > Display architecture details"
@echo "build > Standard build"
- @echo "profile-build > PGO build"
+ @echo "net > Download the default nnue net"
+ @echo "profile-build > Faster build (with profile-guided optimization)"
@echo "strip > Strip executable"
@echo "install > Install executable"
@echo "clean > Clean up"
@echo ""
@echo "Supported archs:"
@echo ""
- @echo "x86-64 > x86 64-bit"
- @echo "x86-64-modern > x86 64-bit with popcnt support"
- @echo "x86-64-bmi2 > x86 64-bit with pext support"
- @echo "x86-32 > x86 32-bit with SSE support"
+ @echo "x86-64-vnni > x86 64-bit with vnni support"
+ @echo "x86-64-avx512 > x86 64-bit with avx512 support"
+ @echo "x86-64-bmi2 > x86 64-bit with bmi2 support"
+ @echo "x86-64-avx2 > x86 64-bit with avx2 support"
+ @echo "x86-64-sse41-popcnt > x86 64-bit with sse41 and popcnt support"
+ @echo "x86-64-modern > common modern CPU, currently x86-64-sse41-popcnt"
+ @echo "x86-64-ssse3 > x86 64-bit with ssse3 support"
+ @echo "x86-64-sse3-popcnt > x86 64-bit with sse3 and popcnt support"
+ @echo "x86-64 > x86 64-bit generic"
+ @echo "x86-32 > x86 32-bit (also enables MMX and SSE)"
@echo "x86-32-old > x86 32-bit fall back for old hardware"
@echo "ppc-64 > PPC 64-bit"
@echo "ppc-32 > PPC 32-bit"
@echo "armv7 > ARMv7 32-bit"
+ @echo "armv8 > ARMv8 64-bit"
+ @echo "apple-silicon > Apple silicon ARM64"
@echo "general-64 > unspecified 64-bit"
@echo "general-32 > unspecified 32-bit"
@echo ""
@echo ""
@echo "Simple examples. If you don't know what to do, you likely want to run: "
@echo ""
- @echo "make build ARCH=x86-64 (This is for 64-bit systems)"
- @echo "make build ARCH=x86-32 (This is for 32-bit systems)"
+ @echo "make -j build ARCH=x86-64 (A portable, slow compile for 64-bit systems)"
+ @echo "make -j build ARCH=x86-32 (A portable, slow compile for 32-bit systems)"
@echo ""
- @echo "Advanced examples, for experienced users: "
+ @echo "Advanced examples, for experienced users looking for performance: "
@echo ""
- @echo "make build ARCH=x86-64 COMP=clang"
- @echo "make profile-build ARCH=x86-64-modern COMP=gcc COMPCXX=g++-4.8"
+ @echo "make help ARCH=x86-64-bmi2"
+ @echo "make -j profile-build ARCH=x86-64-bmi2 COMP=gcc COMPCXX=g++-9.0"
+ @echo "make -j build ARCH=x86-64-ssse3 COMP=clang"
@echo ""
+ifneq ($(empty_arch), yes)
+ @echo "-------------------------------\n"
+ @echo "The selected architecture $(ARCH) will enable the following configuration: "
+ @$(MAKE) ARCH=$(ARCH) COMP=$(COMP) config-sanity
+endif
-.PHONY: help build profile-build strip install clean objclean profileclean help \
+.PHONY: help build profile-build strip install clean net objclean profileclean \
config-sanity icc-profile-use icc-profile-make gcc-profile-use gcc-profile-make \
clang-profile-use clang-profile-make
build: config-sanity
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) all
-profile-build: config-sanity objclean profileclean
+profile-build: net config-sanity objclean profileclean
@echo ""
@echo "Step 1/4. Building instrumented executable ..."
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_make)
#clean all
clean: objclean profileclean
- @rm -f .depend *~ core
+ @rm -f .depend *~ core
+
+net:
+ $(eval nnuenet := $(shell grep EvalFile ucioption.cpp | grep Option | sed 's/.*\(nn-[a-z0-9]\{12\}.nnue\).*/\1/'))
+ @echo "Default net: $(nnuenet)"
+ $(eval nnuedownloadurl := https://tests.stockfishchess.org/api/nn/$(nnuenet))
+ $(eval curl_or_wget := $(shell if hash curl 2>/dev/null; then echo "curl -skL"; elif hash wget 2>/dev/null; then echo "wget -qO-"; fi))
+ @if test -f "$(nnuenet)"; then echo "Already available."; else echo "Downloading $(nnuedownloadurl)"; $(curl_or_wget) $(nnuedownloadurl) > $(nnuenet); fi
+ $(eval shasum_command := $(shell if hash shasum 2>/dev/null; then echo "shasum -a 256 "; elif hash sha256sum 2>/dev/null; then echo "sha256sum "; fi))
+ @if [ "$(nnuenet)" != "nn-"`$(shasum_command) $(nnuenet) | cut -c1-12`".nnue" ]; then echo "Failed download or $(nnuenet) corrupted, please delete!"; exit 1; fi
# clean binaries and objects
objclean:
- @rm -f $(EXE) $(EXE).exe *.o ./syzygy/*.o
+ @rm -f $(EXE) *.o ./syzygy/*.o ./nnue/*.o ./nnue/features/*.o
# clean auxiliary profiling files
profileclean:
@rm -rf profdir
- @rm -f bench.txt *.gcda ./syzygy/*.gcda *.gcno ./syzygy/*.gcno
+ @rm -f bench.txt *.gcda *.gcno ./syzygy/*.gcda ./nnue/*.gcda ./nnue/features/*.gcda *.s
@rm -f stockfish.profdata *.profraw
default:
help
### ==========================================================================
-### Section 5. Private targets
+### Section 5. Private Targets
### ==========================================================================
all: $(EXE) .depend
@echo "prefetch: '$(prefetch)'"
@echo "popcnt: '$(popcnt)'"
@echo "sse: '$(sse)'"
+ @echo "ssse3: '$(ssse3)'"
+ @echo "sse41: '$(sse41)'"
+ @echo "avx2: '$(avx2)'"
@echo "pext: '$(pext)'"
+ @echo "avx512: '$(avx512)'"
+ @echo "vnni: '$(vnni)'"
+ @echo "neon: '$(neon)'"
@echo ""
@echo "Flags:"
@echo "CXX: $(CXX)"
@echo "Testing config sanity. If this fails, try 'make help' ..."
@echo ""
@test "$(debug)" = "yes" || test "$(debug)" = "no"
- @test "$(sanitize)" = "yes" || test "$(sanitize)" = "no"
+ @test "$(sanitize)" = "undefined" || test "$(sanitize)" = "thread" || test "$(sanitize)" = "address" || test "$(sanitize)" = "no"
@test "$(optimize)" = "yes" || test "$(optimize)" = "no"
@test "$(arch)" = "any" || test "$(arch)" = "x86_64" || test "$(arch)" = "i386" || \
- test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" || test "$(arch)" = "armv7"
+ test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" || \
+ test "$(arch)" = "armv7" || test "$(arch)" = "armv8-a" || test "$(arch)" = "arm64"
@test "$(bits)" = "32" || test "$(bits)" = "64"
@test "$(prefetch)" = "yes" || test "$(prefetch)" = "no"
@test "$(popcnt)" = "yes" || test "$(popcnt)" = "no"
@test "$(sse)" = "yes" || test "$(sse)" = "no"
+ @test "$(ssse3)" = "yes" || test "$(ssse3)" = "no"
+ @test "$(sse41)" = "yes" || test "$(sse41)" = "no"
+ @test "$(avx2)" = "yes" || test "$(avx2)" = "no"
@test "$(pext)" = "yes" || test "$(pext)" = "no"
+ @test "$(avx512)" = "yes" || test "$(avx512)" = "no"
+ @test "$(vnni)" = "yes" || test "$(vnni)" = "no"
+ @test "$(neon)" = "yes" || test "$(neon)" = "no"
@test "$(comp)" = "gcc" || test "$(comp)" = "icc" || test "$(comp)" = "mingw" || test "$(comp)" = "clang"
$(EXE): $(OBJS)
- $(CXX) -o $@ $(OBJS) $(LDFLAGS)
+ +$(CXX) -o $@ $(OBJS) $(LDFLAGS)
clang-profile-make:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
all
.depend:
- -@$(CXX) $(DEPENDFLAGS) -MM $(OBJS:.o=.cpp) > $@ 2> /dev/null
+ -@$(CXX) $(DEPENDFLAGS) -MM $(SRCS) > $@ 2> /dev/null
-include .depend
-
projects / stockfish / blobdiff